The spontaneous occurrence of anti-idiotypes associated with the amelioration of disease activity in some autoimmune disorders encourages the view that one may be able to develop a therapeutic strategy based upon manipulation of idiotype networks. Attempts to abrogate autoimmunity by using heterologous anti-idiotype reagents have been rather disappointing and there may well be an expansion of idiotype-negative antibody clones. We argue that idiotypic reagents based on T cells or antibodies derived from the species being treated are more likely to lead to success because they interact more profoundly with the individual's own networks than do heterologous antibodies.